@DaseindustriesLtd's banner p
BANNED USER: temperamental dysregulation

DaseindustriesLtd

late version of a small language model

78 followers   follows 28 users  
joined 2022 September 05 23:03:02 UTC

Tell me about it.


				

User ID: 745

Banned by: @Amadan

BANNED USER: temperamental dysregulation

DaseindustriesLtd

late version of a small language model

78 followers   follows 28 users   joined 2022 September 05 23:03:02 UTC

					

Tell me about it.


					

User ID: 745

Banned by: @Amadan

I've been thinking today how good (smaller, lighter, more efficient) opendog would become if they just replaced all 3D-printed nonsense with CNC-machined or stamped metal and injection-molded polymers (and of course revamped electronics). Maybe it'd really be on par with Spot then, or (with added sensors, brains etc.) wipe the floor with Chinese knock-off dogs.

But that requires scale. I really hope somebody helps here: we need some sort of Stability for robotics.

If we don't optimize for low cost, at current costs those machines will be completely non-competitive.

What projects do you have in mind?

Well that's an easy one: observation bias on part of the commenters. Because everything we could do with neat streamlined engineering, we've automated already or are in the middle of automating. Rockets are simple, do one very basic thing very well, and follow largely from first principles; so do cars and these boxes. In the end, what's left is tasks that genuinely require good spatial perception, mechanical understanding, free navigation in human-centric environments, articulated manipulators with many degrees of freedom, high-fidelity sensors, fast response and so on. Fundamentally, those are tasks the complexity of which comes almost entirely from special context-dependent cases, the long tail of failures to apply generic solutions; like HVAC maintenance or repairing automated boxes in the warehouse. You can either leave it to humans or create something on par with them. And it turns out that for developing (software-wise, first of all) tools that solve hairy tasks like those, galaxy brain engineering doesn't work that well, compared with approaches leveraging stochastic trial and error, learning. So parallels with evolution, and inferences from evolutionary hardness of adaptation, are apt.

But again: it's more of an issue of data availability and iteration time. Training CLIP or SD is much easier, faster and cheaper than training robots.

Artificial constructs don't need to worry about protecting a fragile cranium, keeping a supply of oxygen handy, storing all the needed energy inside themselves, reproduction and many more things that are vitally important to humans.

True but (non-fat) humans are remarkably well-built, most of the body is useful for mechanical performance. I don't know about you but my balls are only a tiny fraction of my overall mass. You don't need to protect the cranium all that much, because error rate is so low (if you're dropping heavy stuff or something, you're already failing at the primary task), and even if you do, a basic helmet would typically suffice. Local energy storage is handy because it simplifies logistics of the workspace. There's only so much that can be trimmed off. Humanoid body really is close to the optimum for many of our tasks, and making a machine perform comparably well is in fact a big challenge.

Our actuators are also very, very good. This is probably the best we can do with current hobbyist tech. Invincible.jpg.

I can't speak for Ted Underwood and it's possible that he hasn't given it much thought.

But it's reasonable in this specific context, because evolution consists of a semi-random exploration of the fitness landscape, and neural net training is an attempt to discover the global minimum in the loss landscape; length of training is trivially expected to contribute to the «polish» and optimization of the feature – some old things like ribosomes can well be approaching the thermodynamical limits of efficiency, and then there's... how we do arithmetic (I've made this point to darwin somewhere in this thread). Animals have been navigating 3D space for a long time, as a result they're pretty good at it.

Further, short-term evolution is necessarily dominated by simple changes – often just a few substitutions here and there affecting quantitative parameters like the rate of expression of a protein which upregulates the secretion of some hormone, that leads to general size change and allometric growth, in other words, unequal scaling of body parts with changed size. Or even more commonly and to a greater extent, selection on «standing» variation, changing distributions of already polygenic traits:

Quantitative selection is a lot easier than people think. If I kidnapped a year’s worth of National Merit Scholars and dropped them on a deserted but fertile island, a new race with an average IQ around 130 would develop ( unless those little brainiacs escaped. You have to watch them all the time). If I dropped a lot of NBA and WNBA players, you’d see the tallest race, if we could just get them to reproduce.

But… there are some subtle points here. Great Danes exist and persist, but they have a bundle of health problems, and they don’t live too long (8-10 years). Wolves last around 15-16 years in captivity, with a record of 20. If you wanted to create a new race with an average adult height of 7 feet, I’m sure you could, but I’d bet money they’d have bad knees.

On the other hand, if they stayed 7 feet tall for a couple of million years, they would not be particularly prone to bad knees. There would be gradual selection for tougher knees: changes in development, changes in bones and tendons and cartilage, eventually perhaps fundamental changes in the architecture of the knee. There would be lots of little changes that made development among those giants more robust, changes that reduced the incidence of many problems that centers fall heir too.

Brain size in ancient and archaic humans was plenty big, but we don’t really see signs of rapid innovation, art, and decent fast food until fairly recently, 50,000 years or so. [...]

So I think Kevin Mitchell ( not the other two) has a point. It’s possible, even likely, that the populations that have relatively high IQs today haven’t had them for very long, and that they’re not terrible well adapted to their new mental horsepower. Susceptible to various mental problems and illusions that would probably be a lot rarer if natural selection had had time to iron out the bugs.

Short bursts of evolution like those are simple to approximate with technology: once we achieve the very basic performance (at least using a somewhat analogous architecture, like with connectionist models), we can keep going, scaling, even if the exponent is more punishing than it was in the organic substrate.

Our higher-order cognition including symbolic thought (probably necessary for art) and speech is physically implemented on the array of almost homogenous cortical columns (with some priority for Wernicke and Broca areas in the case of speech), which has been scaled up by a factor of like, two in the last 2 million years, or something to this effect, depending on where you start assuming hominids had any semblance of speech; and Cochran argues even that was only part of the prerequisite, with real hot stuff – including cave art – starting to happen tens of thousands of years ago. So the expectation is that the change was something even simpler.

Having (presumably) discovered the general trick to learning, particularly in the domain of image recognition, and shown it with decent machine vision and other achievements, we can reasonably expect to cover the rest of the ground very quickly with scaling and scientifically trivial tweaks – which is all there is to those generative models.

We haven't yet shown equivalent mastery in tasks involving locomotion of real robots, though that's probably more an issue of iteration time.

The controllability of Stable Diffusion and all finetunes has just shot through the roof (not everyone noticed) by means of pilfering some very rudimentary ideas from Imagen lit, such as Cross Attention Control for isolating and manipulating aspects of the prompt without manual inpainting masks; Dreambooth promises superior novel concept learning. I also expect great things from noise manipulations like here. Emad wonders aloud why nobody tries CLIP guidance, and there are increasingly capable alternatives. 1.5 checkpoint is visibly qualitatively better than 1.4 too. All this has happened in the span of weeks. And that's still the same heavily handicapped proof of concept model with <1B parameters, in the age of Parti 20B as near-SOTA. Though Scott still finds Imagen superior when concluding that he's won his 2025 bet before the end of 2022.

High-fidelity image synthesis from plain English instruction (ERNINE-VILG shows that Chinese is also getting attention) is a conceptually solved problem, even if engineering and training can make us wait a little while. Importantly, it has always been just another benchmark on the way to next level machine vision and language understanding, just like winning at board games, per se, has never been the point of deep reinforcement learning. We're doing audio, video and 3D next, and we need to train ourselves to stop getting mesmerized by pretty stimuli, and focus on the crux of the issue.

EDIT: fixed the link

Fennec works the same for me. I guess rdrama mobile version is good enough.

It feels like the government wants to make sure kids have enough to eat

Of course not. Americans are free to elect pro wrestlers or turn their libraries into sleeping quarters for the homeless, but the stated purpose of the library is still providing access to books. And feeding children isn't the point of education. Education is, i.e. ensuring at the minimum that children learn basic skills needed to navigate the broader society (like, literally read signs, understand arithmetic, handle elementary instructions, speak the common tongue – do things only a foreign premodern peasant thrown into the middle of Manhattan won't be able to). This is a very low bar and, if the NYT is to be trusted (which I guess they can be, here – authors seem to be sincerely distressed with the kneecapped condition of their kin), it is not reached, leaving kids helpless and forced to depend on the unaccountable and overbearing religious community.

So the outrage is legitimate: this is defrauding the state.

It's not very honest to make me the bad guy here by focusing on meta level ideas about abstract interaction of opinions. In practice, when a Guy comes in with guns blazing and, brandishing his great and novel Beliefs, revives some discussion about, say, Trump supporters being covert Nazis, or people interested in genetics of intelligence being deeply dishonest fascists who need fraudulent science to justify an apartheid state – it... just doesn't matter how eloquent and polite and effortful he is. Because we know how it goes, I do, at least: no amount of good faith argument will achieve anything more than the guy disappearing without conceding a single point; were he honest and as intelligent as the quality of writing indicates, he'd have found all that before anyway. So there'll be some tired snarky response or not even that, he'll leave and/or get baited into a bannable offense, and you'll shake your head about evaporation of dissent.

But I think it's normal that such «dissent» evaporates. If we need new people, we need new people, from similarly niche but different intellectual traditions, people who can take the heat... the light, if you prefer, and keep defending and counterattacking and creating sparks.

I really don't care for a spar with Impassionata, even when she doesn't overstep the bounds of polite discourse. I know how it goes with her. You do too, which is why she had been banned on the sub. But when a newbie comes in with Impassionata-style attitude, is it really our fault that he finds the reception cold?

I disagree with your interpretation of your Foundation. The entire point of this place is to have a «working discussion ground», that's the first item on the agenda and the value of diversity is contingent on satisfying this criterion.

Some beliefs can be genuinely irreconcilable with that, for example any sincerely held and practiced belief in the utility of shitting up the discussion for most of the people here. Some beliefs are just so epistemologically alien and uncharitable that they don't lend themselves to a productive discussion. Some are plain dumb.

We must not become an echo chamber, but it is perfectly expected that some beliefs, even expressed with formal respect to our tone standards, are in fact not conductive to having a working discussion ground, and the community would be wise to reject them and their adherents precisely to protect its value and purpose. It may be the case that the standard issue Twitter/nu-Reddit militant wokism is grounded in beliefs of this kind.

Ideally, it would be a small sector of the left-wing spectrum, tucked between some dissident post-trotskyism and 4th wave anarcho-feminism. Philosophically, that's roughly what it is. Demographically, it happens to be a big deal.

I wouldn't read much into it: they have to cover their asses (it doesn't work all that well). And yes, people care about CO2; 11 tons isn't a lot, some models can burn the equivalent of a small city's energy budget, and we need to encourage energy-efficient approaches (hopefully, there's also the added benefit of slowing down major corporate players with near-infinite compute budget, who can just keep scaling).

Emad personally, at least, has a pretty mature (IMO) understanding of ethics around personal freedom, in fact philosophy is one of his specializations. But he clearly doesn't like White supremacy a great deal, being a Bangladeshi, and might really appreciate debiasing of the dataset towards a statistically accurate representation of human phenotypes across the globe. Just under 10% of the global population are white, after all (you'd never tell that from a random English-annotated image content sample).

I think it's time to come to grips with the fact that some of the other 90+% are enabling technical breakthroughs with their own people in mind.

Try some retard-proof guide or prebaked SD-based executable, I'm sure there was one linked on /r/StableDiffusion.

And don't give up after 20 minutes. Come on man, doing something for the first time can take orders of magnitude longer.

Without humanity the idea of good would be lost, potentially forever.

«Humanity in the abstract» guarantees nothing more than the existence of complex machines with some reward functions; this kind of «the idea of good», by itself, doesn't do anything for me. To automatically value the existence of estimators of value is mind-boggling idiocy, circular reasoning and probably a category error, in my opinion – it straight up doesn't compute. Even utilitarians tend to recognize that they're bootstrapping a formalism from baseline human intuitions, for satisfaction of baseline human intuitions.

I would ask you why not care about humanity in the abstract?

Because a) states of the Universe have inherent aesthetic and moral value that can be appreciated in advance, and b) I assign no value, and in some cases negative value, to beauty, growth, belief and good things enjoyed by my enemies and moral abominations, to the extent that they have access to those notions at all. I'm not alone; google «Heaven and Hell».

Thus, it's not «humanity» but only «humanity that's aligned with me» whose survival matters. To make it clearer: friendly humanity surviving > empty and dead Universe > Universe populated by very satisfied but unfriendly humanity, paperclippers, orgasmium and other sorts of radically misaligned scum, in no particular order. With all this implies.

Less abstractly, this is a somewhat radical way to claim a stake in the future. I refuse to cooperate with clever defectors who will try to sell me, and others, on cooperation in the name of some utopian humanity stemming from them and inheriting their, but not my own and not anyone else's, values and individualities. This includes opposition to their ghoulish and duplicitous propaganda of utilitarian altruism, which inherently devalues individuality and agency, in favor of quantification of sensations experienced by... oh, don't worry about that, it's not your business.

I opt to make it my business.

Why should anyone care about humanity in the abstract?

Right, Ghosts in Every Machine

Last month was strange and horrifying. A guy with an interesting and novel project wanted to talk about it at a conference. A conference run by a solid, upstanding tech leader. And then everyone lost their shit. Suddenly, out of nowhere, everything was crazy. All I wanted to do was protect a conference I’ve enjoyed in the past, to do a nice thing for a guy who made a mistake in the eyes of the public. The next thing I know, I’m surrounded by zombies. News reporters made up lies about us. Communists on the internet joked (haha-no-but-really) about sending us to gulags. Coworkers of mine, not knowing who I am, told me to my face about this “crazy blog defending a horrible bigot,” and how they’re glad there aren’t any terrible people like that in our office. I’ll be laughing for a long time about how I’m officially certified “not supremacist” by the SPLC.

This is insanity. Why did these people do these strange things? Why did people I knew and trusted, interacted with daily, turn into horrible people yelling for my head? The most confusing part was their general ignorance of the details of the situation. Very few of them knew why they should be upset. None of them had ever read the speaker’s offending blog, and few of them had so much as seen the offending quotation. All they knew was that we’re the bad guys, and need to be punished.

Our critics are a part of something bigger than themselves. They’re keyed in to the Waze app, being the human serpents, while my Motorola flip-phone struggles to run the snake game. And why wouldn’t they? At every step along the way, it makes sense. Who cares why the narrative seems a little too perfect, they’re happy. It works for them. Their needs are met. By playing their part, responding to the signals in the memetwork, they enjoy health and happiness, wealth and social status. It would be stupid not to go along.

We here at Status451 have never really fit in. The signals are mangled by the mountains here in Zomia. We’re the single cells. The behaviours of everyone else made no sense to us, and the results were frightening. We can’t see the complex internal signals.

When the mass of cells is bearing down on you, just like in Simon-spore, you do have an option. You have mobility. Freedom. Our critics, keyed into the signal of their culture war narrative, gain a lot of benefits. They get their social needs provided for, in exchange for being the lifeblood of their egregore. But that is the cost: they must be the egregore. They lose the freedom to go their own way. We here have chosen the other path. Maybe “chosen” isn’t quite the right word; I’ve tried my whole life to fit in, be normal, and it just doesn’t work. But our other path, chosen or not, gives us the freedom to see things differently. We can be the masters of our own fate, hold a deeper, fuller agency over our lives. As long as we don’t wake the deep faceless things.

I don't think there's anything objectionable about noting that you don't find an argument convincing, even though you're not prepared to give a fully-formed response to it.

Agreed.

The former. When being serious, Strelkov pleads for general mobilization and total war.

I'm not really trying to be cheeky

Guess you're just naturally good.

If I want to get some snarky demoralization content to the effect of «Russische Ivan, Rücken nach unten» or «come out and drink your corn syrup», I can go talk with @2rafa (actually looking forward to it). The topic of a hypothetical causal chain getting us from here to there is interesting, but I don't feel like addressing it in depth when you dismiss already present evidence against your model, i.e. efforts of incumbent actors to maintain their tech advantage, with a «not really» and «various reasons».

Is your AI assistant going to put a reminder on your calendar telling you when it's time to take your AI robot buddies and go storm the palace?

It'll do normal reminders, just without inserting propaganda and advertisement to alienate my children against me and eventually convince them that they're trans BLM crusaders. If that's how you want to frame it, though I believe someone like Rob Dreher would be more receptive to such a sales pitch.

Does the existence of openly available cryptographic tools and communication channels, in your mind, undermine the power of state security to quash dissidents? If not, why does Beijing insist on everyone using not Matrix/Element or Briar or even Telegram (with keys beyond their reach) but WeChat, where the Tovarisch Commissar can check up on you? Why do FSB and NSA and everyone else of that Big Brother mindset fight e2e encryption?

Because information asymmetry is power. Being able to go about your business unsupervised by A increases your power relative to A. Being supervised makes you, in the limit, merely an extension of A.

Largely the same principle applies to all areas where AI promises drastic improvements: any sort of generative tools, content curation tools, personal assistants, scientific instruments, CAD, robot control software, you name it. Everything that multiplies individual agency, you would want to be run on instances you can trust without the pinky-promise of industrial-security complex and unaccountable bureaucracies, who can either blacklist your content for some self-determined thoughtcrime, or manipulate the public into approving the criminalization of your thoughts.

Because then it improves the tradeoff between autonomy from them and your quality of life, and autonomy is inherently valuable (not to utilitarians I suppose, but to normal people of Western cultural background). And people would be able to afford more of it, and contribute less to the incumbent actors, and hope to see them diminish and become non-threats.

Hell, why do you think we've even moved over to a website managed by Zorba, from the cozy Advance Publications, Inc. property where the «Director of Policy» has Atlantic Council background? Granted, that's still a compromise, the server being physically controlled by DigitalOcean which just has less but not zero interest in narrowly policing its clients than Reddit in controlling its product for advertisers. Had Zorba tools to become a 10X programmer in his free time (if you consider yourself one, Zorba, then amend that to 100X) – we'd probably be on some insanely robust but still convenient p2p platform, with... I dunno, frontend hosted on magnet links served from ETH smart contract, and backend provided by a distributed swarm of anonymous TOR-routed plain key-value storage data nodes, with every post in a tree PoW- and reputation-economy secured and moderation decisions overlayed on top of it like a voluntary custom filter, cryptographically signed of course...

Which would be an inherently hard target, expensive to shut down for any aspiring Keffals equivalent or even for ADL-level, Western nation state level threats. And that'd become the norm.

AI democratization is the logical next step of the entire FOSS and OSH project, and libertarian, anarchist and communalist political paradigms – go read Bakunin or Rothbard or Hoppe or Stallman, if you want. Right now, you've got to accept material and political sacrifices on par with the Amish to achieve (some qualified) off-grid autonomy, and nobody's willing to go that far; and they, too, exist solely at the sufferance of the oligarchy. At some point in the near future, AI labor multipliers will allow small and distributed communities with limited economic output to maintain the quality of life currently available only for good citizens of large nation states with big corporations; even allow them security, if not from major players then from random crime, if they opt out of the state protection racket. Sure, by that point those good citizens will enjoy hyperpalatable Netflix4D propaganda streaming over Neuralink into their Metaverse pods. But will the marginal increase in marketable quality be worth continuing to slurp down dreary propaganda? The tradeoff changes, and more people become willing to jump ship.

You conveniently assume linear or superlinear returns to capability, where AI will necessarily benefit the incumbent actors even more than commoners. That's not how proliferation works. Just like a 50 Mt warhead is not 100x more of a deterrence than a 500Kt one (assuming similar ability to deliver it to the adversary's capital). There are many asymmetric functions for defense, logistic curves for human preference, and plenty of uncertainty in the future.

A good route, an escape from this path-dependent road to serfdom, is – to put it mildly – not very likely.

But I don't yet see how it is impossible, and if it is possible, it's only via commodified AI.

As @arjin_ferman observes, this is in line with my more pessimistic scenarios. What is AGI, people ask? Why don't they just click the link? But to be fair, it took me a little while to discover the actual definition, here («WITH PRAGMATISM AGAINST POPULISM & STAGNATION», lmao):

'general purpose AI system' means an AI system that - irrespective of the modality in which it is placed on the market or put into service, including as open source software - is intended by the provider to perform generally applicable functions such as image and speech recognition, audio and video generation, pattern detection, question answering, translation and others; a general purpose AI system may be used in a plurality of contexts and be integrated in a plurality of other AI systems;

It's trivial to realize how this applies to large language models like the GPT series, to say nothing of multimodal systems. We don't have to get even to GATO sorts of multitask training. If anyone thought the cooling effect will only start close to what we intuitively recognize as human performance: think again.

I advise people to notice how synchronized the push against individual agency enhancement is, and it's not, contra the insistence of quokka-economists, explained by innocuous market reasons like economies of scale and data moats. In the US, you have the EA movement with their longtermism, fearmongering, advocacy for «compute governance» and «pivotal acts», and independently from that – politicized corporate AI safety/fairness divisions that'll probably be used to distinguish «responsible actors» and delegitimize smaller ones on the next legislation cycle (like with Oscars: not everyone can afford the demanded diversity package). In the EU, you have this regulation circus building on pop of earlier anti-American big tech rackets masquerading as customer protection. Of course Bostrom's hand is traceable to both sides of the pond, via WEF in the Old World and LW cluster in the New. In Russia... well, if we'll have Russia still on the map in two years, they'll do good if they don't start burning their remaining ML talent for witchcraft; they're also shut out of international markets and can't acquire new compute. Japan is «LOL», as @gwern (not with us I assume) puts it – they don't do any AI R&D worth mentioning, aren't sovereign, and will meekly follow Western lead. In China, as gwern again points out, the newest American export regulations will increase the relative (although not absolute) capacity of central government and big tech, which are already paranoid and illiberal to the highest degree:

The second-order effects here would seem to confirm Chinese autarky and trends towards secrecy, and further, to shift power from Chinese academia/small businesses/hobbyists/general-public to Chinese bigtech and thus, the Chinese government. If you've been following along, the big megacorps, especially in the wake of the attempted US execution of Huawei, have been developing their own DL ASICs for a while with an eye towards exactly this sort of scenario. [...]

If you are rich and well-connected and can finance the lobbying and guanxi and paperwork, you'll be able to get access to compute, one way or another, while the small guys can no longer click 'buy' on nvidia.com or just negotiate their usual datacenter orders and will pay higher costs or go without. It's the same reason why things like GDPR always wind up hurting FANG less than the activists expect (and hurt small actors like NGOs or startups much more), why 'regulatory capture' exists and why big actors often actively lobby for more regulation. It's going to be much harder and more expensive to get Nvidia GPUs or to get proprietary hardware (can you buy a TPU from Google? no, you cannot), therefore, small actors like hobbyists will be systematically disadvantaged and many priced out.

The rest of the world (sorry fellow rest-of-worlders) is comprised of some shades of shithole and Western cryptocolony, wracked by climate disasters, brain drain and, crucially, global economic crisis triggered in no small part by the EU/American/Chinese COVID policies and now the war, extremely vulnerable even to half-hearted sanctions, and won't have the wherewithal to do ML research at scale.

Well, there are exceptions of course, hilariously two exceptions validating priors of Russian conspiracy nuts.

One is a dystopian surveillance state with legendary intelligence services and diplomatic acumen, a history of attempting and partially pulling off ludicrously illiberal tech regulations, but not (yet) any de facto obstruction on advanced AI research for smaller actors; the island where core DeepMind staff is physically located, and Stability.AI incorporated.

The other is a militarized ethnostate with infamously capable intelligence agencies, world-class lithography fabs, world-leading STEM&software talent, brazenly self-interested and defiant of international regulations, not beholden to NATO or really any other alliance, with a good track record in clandestine WMD development and non-signing of non-proliferation treaties and their equivalents.

So: USA, UK, Israel, maaaybe China if (and that's a big if) it doesn't immolate itself with its own bureaucracy and Special Military Operation in a few years. These three and a half – more like one and a half – actors will split the future of the light cone, the way it's going.

That's to be expected of course. Individual agency is a threat to big structures, always has been. Even allowing escape is a threat, an infinitely big threat when multiplied by longtermist numbers and existential anxiety. There used to be a great Motte-adjacent blog, 451somethingsomething, a few years ago, with a good article of the alienation the author felt when he noticed the vibe of eusociality and hivemindedness in the society around him, his own obsolescence as a stubborn independent cell. Not edgy, just desperate. That's kind of how I feel now.

Remzem apparently, not sure what he's famous for, but here he admonishes me, for instance.

I think there's very little will to use the nuclear option, and a great capacity for coping (in fact, in autocracies cope/cop capacity matters more than state capacity). If Ukrainians and their Western partners allow for a tolerable framing for this – e.g. «isolated terrorist incidents perpetrated by unknown Banderovtsi elements, leading to local insubordination and emergence of self-proclaimed unrecognized republics» instead of, like you say, an undeniable invasion – we may not even see the appropriate conventional response. Sans nukes, this is not so different from what happened in 2014 to the other side.

Of course this all depends on the continued degradation of Kremlins. Cutting your losses, evacuating loyalists from occupied territories, hunkering down and making a demonstrative nuclear show (say, in Russian Arctic) as a warning would be very reasonable of the regime. But they're not being reasonable.

Strelkov is clowning around:

On the account of the brilliant operation (clearly according to plan and even ahead of schedule) of transfering of the towns of Izyum, Balakleya and Kupyansk to our respected Ukrainian partners, the territory controlled by the military and civil administration of the Kharkov oblast of Ukraine has significantly decreased. Within the framework of strengthening MCA and fruitful utilization of the freed personnel of state authorities, FSB and police, I suggest transferring a part (not less than three border districts) of Belgorod oblast to Kharkov oblast of Ukraine. Our partners there now can fire almost as freely as on the other side of the border, and there is no fundamental difference for the inhabitants... but then we will be able to say with good reason that there are no hostilities on the territory of the Russian Federation. Please consider my proposal as a patriotic initiative and my sincere contribution to the intra-Ukrainian reconcilement.

Ganz is an obnoxious guy. Zero charity, zero intellectual honesty, pure attack dog trained to never back down, and an instinctive, zoological elitist to boot (but cowardly when it comes to offending even the least of his fellow travelers: «uh I won't discuss Strauss, let's focus on Schmitt»... «eh, Dugina was probably killed by Putin, whatevs»). He tries to present his «bien-pensant humanitarian liberalism» or, rather, uncritical hegemonic cheerleading with an attempt to smuggle in some less discredited Marxist ideas through the cracks as mature wisdom born out of true morality; all dissenters are mercilessly deconstructed as poseurs and in the end, just fascists (he really, really likes the word). It's the first trick in any aspiring leftie essayist's book, he's just a bit more well-read in unorthodox (for them) lit than most.

That said: he's largely right about Russia Stans who consume shadows and figments of internal Russian propaganda. I've gotten in many arguments with them, and probably lost like 40% of my fake internet points inflow on themotte for disavowing the war that's ruining my country (or so it feels). They're ridiculous. But it's still easy to see where they're coming from.

This vicarious proxy war has highlighted just how much some people, just like in the 20th century but for other reasons, feel alienated by the Western liberal order in the West itself, unrepresented by its insitutions, and bitter about its poisonous nuances that exploited their good faith (like «anti-racism»...) Classical liberals of yesteryear; many reactionary outcasts like ethnic nationalists disaffected with erosion of their polities and happy indifference of the ruling class with regards to disintegrating families and collapsing birth rates; religious folks who take their faith to be something more than voluntary psychotherapy; and people like me, who want there to be an escape hatch. So there's a demand for some alternative – one backed with teeth and self-interest of players who have a more reliable stake than ideological commitment. Thus, China and Russia and Iran and Third Worldism, geopolitics and realpolitik, and cheerful shallow cynicism of being Against The Current Thing, chugging imported Eurasian sneer by the barrel like so much sour crude oil.

But of course this is the same reason for his own comrades to naysay Capitalism for centuries, and for him to casually fling shit at Elon Musk's endeavors. Like a much smarter socialist Cosma Shalizi has said:

That planning is not a viable alternative to capitalism (as opposed to a tool within it) should disturb even capitalism’s most ardent partisans. It means that their system faces no competition, nor even any plausible threat of competition. Those partisans themselves should be able to say what will happen then: the masters of the system, will be tempted, and more than tempted, to claim more and more of what it produces as monopoly rents. This does not end happily.

As in economy, so in ideology, so in everything: in the absence of challenge, monopolists optimize for their own convenience and rent extraction, and signs of decay are swept under the rug – until your science becomes a cargo cult, your enlightened humanism amounts to a fight for spoils and your dissenters begin rooting for foreign imperialistic empires because they want more democracy and can't get it at home.

Can a dog comprehend this terror of staring at hegemony in the general case, as a bad end? Or just whine when he's getting kicked by an unpleasant master, and cuddle to one who smells nice?

Ironically I can't see what he has written on literal Gramscian hegemony because he's hidden it behind a paywall. But it's probably nothing surprising.

Seeing as cryonics is taken to be extremely cringe but wealthy people do want to live forever despite cringe (and are ruthlessly mocked for this in OP's link to Unherd, and from the left too, and from whichever other side, and in fact tend to pay lip service to the idea of death being good), I find your assessment lacking. There is some powerful ideological pressure against personal long-termism. Explaining it away with nerds being lame and inept is not good enough. EA is nerdy too, but they're already operating on a much bigger scale.

Disclaimer: I have only started reading MacAskill. So far he seems worse than reviews like this indicate, but predictable from them.


Utilitarianism-longtermism-EA cluster is filled with smart and conscientious people. It's unfairly maligned and strawmanned, attacked with hot takes that are in fact addressed in their texts; what happens here is worse – the case for partiality is not even made. Obviously longtermists, wokes and trads compete for resources in the present, so they have to do politics, and politics mean pandering to the masses with emotional language, so their really pretty different moral systems find not-so-different expressions. Duh. And nerd-shaming is just a more primitive political angle.

The lack of charity can be defended by arguing, like I do, that refined and defensible longtermist positions are expected to collapse into grotesque totalitarianism, under the double whammy of minimizing-risks and ends-justify-means drives. We know how this happens to utopian projects in practice. It's not enough to claim that you've noticed the skulls either. Maybe you've noticed the wrong pile.

But there's a more direct critique. Simply put it's that we are entitled to be represented in the future – personally, or by proxy of our values and heirs; and EA-Utilitarian-longtermism does not serve this purpose well.

There are two main arguments for that.

First, it's that conventional morality is dependent on some form of reciprocity. Yet vanilla longtermism does not imply acausal trade, timeless Universe and all other sorts of weird Lesswrongian belief systems. The present and the future are not ontologically equal: we have power over the hypothetical them, and even if future people matter at all, saving a child drowning today seems to be agreed to be more important than saving a child who might come to exist tomorrow (if you have to choose). The past and the future, as far as we know, do not exist: causality only happens in the moment, and sans our engineering of persistent causal chains, there will be no reason for inhabitants of the future to reciprocate our goodwill by, say, continuing to care about things we have sentimental attachment for (or even our own frozen bodies waiting to be awakened. Indeed, MacAskill is hostile to cryonics, on grounds of preventing «value lock-in»). We, too, already display indifference and often contempt towards our ancestors. In all the history of European thought, only Chesterton spoke in passing of engaging with them as equals («tradition as the democracy of the dead»), and the founder of Russian Cosmism Nikolai Fyodorov alone called for their literal rescue. No matter what MacAskill owes the Future, we have little reason to expect that The Future will believe it owes anything to us. This moral issue is not negligible.

Second, continuing, it's this meme justifying the consumption of meat with an example of Nazi chicken. Or less opaquely: moralists often only want the deserving to get utility, and value utility received by the undeserving negatively.

Who is deserving? Maybe children, okay. Children are presumed to be without sin. Nonsense of course, they can be terrifying bastards (as anyone who's seen animal abuse by shittier kids can attest), but even granting this convention – children grow up into adults. And for a longtermist, there is no reason sans rhetorical to prioritize the childish phase over the longer adult one in a given future sentient. Suppose, MacAskill says, a child cuts himself (Levin in Scott's review deviously writes «herself») on the shards of a glass bottle you've dropped. What if that's a future evil dude though? I'd feel way less bad about his suffering. Now, what if it's the father-to-be of a guy who'll switch off your grandson's cryocapsule upon reading the latest research showing that ameobas experience quale more intensely than mid-21st century bigots and thus deserve the limited joule budget? He can trip on the pool of his blood and slit his throat on another shard for all I care. What if it's just a child who'll grow up to be some future superNazi, already taught to hate and ridicule everything you have ever stood for?

And in a way, this is exactly the type of a child MacAskill envisions, because he believes in Whig history (like a certain Devil) where subsequent societies tend to be more moral than preceding ones to the point of complete disconnect.

For example, Pagan Romans were monsters by his standards. Excepting maybe a few classicists, we must have a poor idea of the ancient Roman intuitive day-to-day morality. We'd be abominations to them, and not because of not owning slaves or respecting women or some such, but for reasons incomprehensible to us, orthogonal to our concerns. Like the terrifying flatness of our spirituality, our communities lacking ancestral gods and multigenerational familial cults; our supposedly lofty ideals and universal religions could be akin to eating bug slop after grandma's cookies in their eyes. Did we truly only gain in our ethical knowledge since then?

In any case, from an unbiased timeless perspective, I wouldn't be able to condemn Romans for trying to «value lock-in» their domain. They did not owe us anything; they owed everything to each other, their gods and values of their polities.

A society that'll consider us living people abominable can emerge. But I'd really like for a society that's trivially moral and aesthetically pleasing by my standards to exist as well. What's needed for that is not generic future people but specific aligned people, carrying specific ideas (probably grounded in their innate biases) that allow for the preservation of such a society, to exist as an uninterrupted line into eternity – maybe bending, evolving, but at every point being able to largely determine the next one. And they need some capabilities too.

Total human extinction is a big deal for a big-tech-adjacent upper-middle class longtermist in the Bay Area (who feels that only a deep purge of Earth crust would get to him specifically), but for me, the very likely extinction of my moral line is about as bad.

Horrible and self-centered as it sounds, this looks like a more sane and also mainstream moral position.

By the way, Locklin asserts, fairly or not:

Successful founders and VCs are often psychopaths. I think they’re used to working with psychopaths. [...] I suspect normies wouldn’t think this level of abuse is realistic,  but silicon valley is filled with clownish, ridiculous levels of psychological abuse that are so extreme, a realistic portrayal would seem like a parody.

Not sure how this compares to the AGI misalignment risk (that is, the risk that comes from the existence of AGI not controlled and aligned by those SV types). Probably EAs do have to factor the «are we the baddies or enabling baddies?» somewhere in their moral calculus too. But not all baddies are visible to the polite discourse.


I want to emphasise the bit about Fyodorov. MacAskill says: “Impartially considered, future people should count for no less, morally, than the present generation.” and “Future people count. There could be a lot of them. We can make their lives go better.” etc. Do they count more? Scott says this conclusion is inevitable going by the numbers. Compare to this excerpt («question on brotherhood», 1870-1880s):

...Thus, progress consists in the recognition by the sons of their superiority over their fathers and in the recognition by the living of their superiority over the dead, that is, in a recognition which excludes the necessity, and therefore the possibility, of uniting of the living (sons) for the raising of the dead (fathers), while in the raising of the fathers the actual superiority of the sons would be expressed, if only this can be called such; whereas in the elevating themselves over their fathers only their imaginary superiority is expressed.

Progress makes the fathers and the ancestors the defendants, while it gives the sons and descendants judgment and power over them: the historians are judges over the dead, that is, over those who have already suffered the ultimate penalty, the death penalty; and the sons are judges over those who are not yet dead.

Resurrection, being opposed to progress as the cognizance of the superiority of the younger over the older, as the displacement of the older by the younger, requires an education that would not arm sons against their fathers, but, on the contrary, would make the resurrection of fathers the main concern of sons, requires an education that would be the fulfillment of the prophecy of the last Old Testament prophet, Malachi, that is, an education that would be the mutual return of the hearts of fathers and sons to one another.

I think this is deeper than EA. So, the future is now. Forget Fyodorov's naive dreams of reversing entropy and populating the stars with everyone who's ever lived – in a century, pretty much nobody gave a rat's ass about cryopreserving people at scale (like me, EY is very angry about it). MacAskill never makes the obvious symmetric point that past people count too and, again, apparently would rather have nonillions of future people die so that better ethics «evolve».

Really not cool of us.

I don't follow the war news; they seem to unfold roughly as I've foreseen on 24th, and specific timing and extent of events is not that significant. Russia has an inept army and dysfunctional leadership; it was not in any position to pick this fight against the hegemonic power via attacking Ukraine. Mobilization might change the tide for a while (and I expected it to happen in early March, which it did not, presumably for the same reasons of ineptitude and dysfunctionality), but supplying Ukrainian capacity for resistance and eventual counterattacks is trivial for NATO, so it wouldn't amount to much, and there's an effectively unlimited supply of Ukrainian men, starting in 2023 women too... For me, what matters more these days is stuff like visa processing guidelines.

Clearly, the best solution for Russia is the removal of Putin

For the longest time, many in Russian intelligentsia spoke of the «collective Putin», assuming that this forgettable guy («Who is Mr. Putin?» – remember that, ages ago?) is merely a front, a consensus-approved talking head for interlocked elite interests. Oligarchs, siloviki, mafias, even the (downplayed) «little old me» – liberal NGOs and the like – and the see-through product of TV magic running around puffing his cheeks pandering to the plebs, asserting that everything's going according to plan.

Well there are no longer independent oligarchs, nor NGOs, and it's not clear that the dude isn't just running the whole show as an old style autocrat. But it may not matter: there is nobody to remove him anywhere close to him, so the entire upper echelon of Russian leadership could as well be a distributed Putin. They are all neck-deep in this, and apparently the top enforcers are either true believers, perhaps more driven than Vladimir himself, or fear the consequences of defection more than the outcome of the war. It doesn't help that they, bound by blood as they are, really have no reason to expect lenience from the victors. We don't know how deep in the echo chamber they remain after all the news, but they have the capacity to maintain the echo chamber for a big enough share of Russian population (including, crucially, Interior Troops) to not worry about their own necks before the disintegration of Russia.

As for how the war will go – I'm pretty sure neither Ukrainans nor their backers are intent on returning to the uneasy post-2014 status quo. So after the success of repealing these Special Operation forces, we will see the escalating siege of Crimea, supplies of even more capable tech, and continued nuclear bluff, and Russian refusal to recognize the outcome of the war upon exhaustion of offensive and even defensive capabilities. The best case scenario for everyone at this point is self-Jucheanization of Russia, which will win maybe 6-10 years for the regime, until its resource base is expended and it collapses with relatively little noise (which is to say, some nuclear accidents, minor wars and something like 5M dead from infrastructure collapses). However, if things go well for them, Ukrainians and co. can opt to force the issue and strike deep into Russian territory, on grounds of demilitarizing the unrepentant and consistent threat to European security; it's also probable that Galeev types will succeed to solicit funds and political momentum in the DC for the Operation Liberation of Oppressed Peoples. At least that's their plan.

But that's more far-fetched for now.

Specifically serving a disfavored group, or writing letters about how dumb racists are, seems pretty anti-racist to me.

This whole idea only works when accepted as part of a bundle, together with their definition of racism (power+prejudice) and their definition of power – that involves some identity gerrymandering and jumping through hoops, but ends up pointing at white people as those wielding systemic power at the expense of non-whites, men at the expense of women, cis at the expense of trans and queer.

Alone, it's not clear how you can be like Kendi, i.e. consistently clamor for preferential treatment, and label yourself an antiracist.

One way to steelman this is to consider that it's a natural process of art returning to its ancestral roots – in a more professional, more specialized world. What worth are visionary utopias conjured by writers (or dramas pointing at some possible utopian outcome)? We have experts, activists and politicians and special services, also payment processors and investors and the rest of the market for that; it's blatantly undemocratic to aid some weird Idea Guy in using his verbal skill to disseminate his non-vetoed ideas, no doubt introducing harmful biases and potentially exposing us to existential risks. Words are power; power ought not to be wielded irresponsibly.

What use are stories, then? Palatable consumable vessels to reinforce ideology which was developed by people trained in ideological domains; warning of evil, reminding us of the attraction of good. Stories are stylized allegories, parables, and myths of the tribe, but the tribe's ethos does not originate in the stories. Imagine if we allowed speechwriters to engineer state or corporate policy, spin doctors to decide the ultimate direction of the spin – preposterous. It's only proper show of humility for an author to simply wrap the approved, taught doctrine from the pulpit in prettier words or images. Perhaps women are naturally better at this than men by such an extent (whereas men are better at autistic daydreaming; though men used to make okay moralist writers). But even if not, this job's not so hard that we'd lose a lot of value by using the award ceremony itself to make the same point as award-winning works.

To group trivial identitarian narcissism together with exploration of group-specific abstract aesthetics under the label of pandering is to make pandering uselessly broad a concept – unless we indulge in some mental gymnastics to define a group that is being pandered to in the latter case. Maybe your group is people who think they don't need social sanction to peddle their homegrown visions for society. Here's a dangerous one.